diff --git a/.gitea/workflows/deploy.yaml b/.gitea/workflows/deploy.yaml
index 79673e44..d9de5145 100644
--- a/.gitea/workflows/deploy.yaml
+++ b/.gitea/workflows/deploy.yaml
@@ -10,8 +10,23 @@ jobs:
       - name: Run Deploy Scripts Locally
         run: |
           set -e
+
+          # Define a lockfile to prevent double, concurrent deployments
+          exec 9>/var/tmp/epic_web_control_deploy.lock
+          flock -n 9 || { echo "ERROR: Another deployment is already running! Cancelling."; exit 1; }
+
           echo "--- EPIC WEB CONTROL: Starting Auto-Cleanup & Deploy ---"
 
+          # Fallback routine: If anything crashes during the steps below,
+          # try to keep the current service running so the site doesn't stay down.
+          error_handler() {
+            echo "!!! DEPLOYMENT FAILED on line $1 !!!" >&2
+            echo "Attempting to keep the current service running..." >&2
+            sudo systemctl start atom-nexst.service || true
+            exit 1
+          }
+          trap 'error_handler $LINENO' ERR
+
           # 1. Clean up unused build images safely
           docker image prune -f
 
@@ -29,14 +44,18 @@ jobs:
           git reset --hard origin/main
 
           # Preserve .next/cache so Next.js can reuse its incremental build cache.
-          # next build safely refreshes stale server/static output itself.
           rm -rf .output dist
 
           # 4. Configure trusted dependencies globally and install cleanly
           export PNPM_CONFIG_ONLY_BUILT_DEPENDENCIES="@parcel/watcher,@swc/core,sharp"
           pnpm install --frozen-lockfile
 
-          # 5. Apply versioned CMS migrations and generate the Prisma client safely
+          # 5. [SAFETY] Database Backup & Migration
+          # TIP: Adjust this to your database (e.g., pg_dump for Postgres or cp for SQLite)
+          echo "Creating database backup..."
+          # Example for SQLite:
+          # cp prisma/dev.db prisma/backup_prior_to_deploy.db
+          
           pnpm db:migrate
           pnpm prisma:generate
 
@@ -49,9 +68,10 @@ jobs:
           # 8. Hard restart of the Systemd service to clear memory cache
           echo "Hard resetting systemd service..."
           sudo systemctl stop atom-nexst.service || true
-          # Kill any lingering next-server processes holding port 3000 so the
-          # new instance can bind cleanly (systemctl stop may leave orphans).
+
+          # Kill any lingering next-server processes holding port 3000
           pkill -f 'next-server' || true
+          
           sudo systemctl start atom-nexst.service
 
           # Extra health check: Ensure the service is actually running
@@ -61,4 +81,4 @@ jobs:
             exit 1
           fi
 
-          echo "--- Deployment successfully completed ---"
+          echo "--- Deployment successfully completed ---"
\ No newline at end of file